Comparative Study of Regression Models Applied to the Prediction of Energy Generated by a Micro Wind Turbine

Abstract

To create an efficient electricity system that includes renewable energy generation, able to respond correctly to the variability of power generation, it is essential to establish future production. For this reason, in this work, the performance of different regression models when predicting the energy produced by a small wind turbine based on meteorological variables is compared, seeking the best predictions. Four methods are evaluated: polynomial, bayesian, support vector machine, and artificial neural network. Several metrics are used to compare the models, such as Mean Absolute Error, Root Mean Squared Error, Mean Squared Error, Median Absolute Error and Coefficient of Determination, along with hypothesis testing.